1. Field of the Invention
The present invention relates to a medium scan method for monitoring information written to a medium surface in a preventive and maintainable manner in a disk device which executes a read/write process of information to/from a medium surface by request from an upper device such as a host computer or the like.
2. Description of the Related Art
There may be cases where a disk device such as a magnetic disk device, which writes/reads to/from a medium surface data transmitted from an upper device such as a host computer or the like and transmits the data to the upper device, cannot maintain the original quality of the data because of a disturbance (such as the attenuation of a data signal due to an influence of leakage flux at the time of a write to a contiguous region, a partial crush of contiguous data due to device vibrations, or the like) or a change by a time elapsing.
For example, there is a problem such that if an unrestorable data check or a restorable data check, and a sector replacement process accompanying the check occur when attempts are made to actually read data from a sector to which data is successfully written, an upper device comes under a significant influence such as a decrease in a processing speed or the like.
Such problems have become obvious as the recording density (BPI: a density in a recording direction (circumferential direction), and TPI: a density between recording tracks (in a diameter direction)) of a medium surface improves with a recent increase in the storage capacity of a disk device, leading to a large obstacle to the quality of the disk device.
Therefore, a function for reading/verifying the entire surface of a medium at predetermined time intervals while a disk device is idle has been popularized as a function for assuring data on the medium surface beforehand.
Namely, the medium surface is internally read while the disk device is idle, and (1) if a sector which causes an unrestorable read error is detected, it is recorded as log data, or (2) data is rewritten to a sector which causes a restorable read error.
As a result, a temporary read error or head noise, an influence exerted by leakage flux at the time of a write to a contiguous track, partial damage caused by a vibrated write can be possibly relieved. Additionally, a restorable sector can be prevented from being degraded to unrestorable state before the degradation occurs.
Furthermore, (3) if an error occurs even in a read remade after the rewrite in (2), this can be a true fault of the medium. In this case, a sector replacement process is executed if possible.
The above described processes are generally called background medium scan, etc., and its specifications are actually proposed by ANSI (American National Standard Institute).
However, the conventional background medium scan (hereinafter referred to as medium scan) is made for the entire recording region of a disk device at predetermined time intervals. Therefore, it cannot flexibly cope with read/write processes executed by diverse requests from an upper device. For example, the following cases are considered.
(1) Case Where Writes Concentrate on a Specific Cylinder Head Sector
To reduce ill effects such as side erase caused by writes which frequently occur to a specific region, read scan for a contiguous region must be started at a time when definite writes concentrate on a definite region. However, with the conventional medium scan, an entire region is always targeted to be scanned without being dependent on an access pattern.
(2) Case Where Writes to a Specific Cylinder Head Sector Occur in a Short Time
If writes frequently occur to a specific region in a short time, scan must be started at a time of exceeding a specific threshold time. However, the conventional medium scan is always made at predetermined time intervals without being dependent on the time of exceeding a threshold.
Japanese Patent Publication No. SHO58-182775 discloses a self-diagnosis method for use in a magnetic disk device, which can check an arbitrary region on a magnetic disk in an idle state.
Additionally, Japanese Patent Publication No. HEI04-205972 discloses a magnetic disk device which can make a reliable diagnosis in a short time.
Furthermore, Japanese Patent Publication No. 2005-190003 discloses a hard disk drive, an information processing device, a disk controlling device, and a diagnosis method for a hard disk drive.
Still further, Japanese Patent Publication No. 2005-322399 discloses an error state processing method and device for restoring a track squeeze fault, etc. of a magnetic disk drive used in a storage device system of large capacity.